Well you guys, I promised I'd make this thread a while back (during the MGC if I recall), so here you go.

This thread is meant for those of us who have already been through all or part of a homebrew pinball project, so we can all share our experiences and pitfalls for the rest of the community. I'll be adding more posts below as I think of things I dealt with in the process of building Undertale, and with the impressive stuff I've seen from the rest of you I'm sure this can turn into a very nice resource for future homebrewers.

Keep in mind that every pinball machine is different, and every homebrew is unique in both design and construction choices. Some of the advice offered in this thread might not apply to you based on your decisions. For instance, most people are probably not crazy enough to go and build the entire cabinet, electronics, and software. If you do, however, much of my own experience might be of use to you. Mental help might be of use to you as well in that particular case!

Since my project started from scratch, I will do the same on my posts for this thread. If you don't intend to go this extreme, focus on other people's posts.

Rule number one. TAKE YOUR TIME.

When you build a custom game, you're taking on a personal project. You're not on a professional team with a looming deadline. The only customer is you, and the way to please that customer is to build something you can be proud of. Don't rush your reference measurements (if you're copying an existing cabinet design). Try not to get antsy and assemble your cabinet before you've remembered to cut and route every necessary hole. If you have other activities going on, don't try to cram design or assembly in between them - wait until you really do have some free time.

This is what happens when you try to rush cabinet construction. You get horrific squaring disasters like this. Don't be like this person.
IMG_20180607_221313262_LL (resized).jpg

Don't be afraid to go back and cut just a little bit more off an edge, or even (after a few deep breaths) cut a new panel if a hole is misaligned or botched beyond repair. If, after all the work you've done, your playfield sits 1/4" lower in the rear left corner than on the other side... you're going to be sad.

As a general rule, your lower cabinet needs cuts for the following components. Make sure these are all addressed or else you'll be filling an assembled cabinet with sawdust as you haphazardly cut the necessary holes with drills and saws at awkward angles. No, I definitely didn't do that at any point...
-Lockdown bar reciever, mounted with carriage bolts. The method that I used was to assemble the full cabinet, unpainted, dragging the carriage bolts all the way in and letting them create their square holes in the process. I then disassembled the cabinet, painted and dressed it as desired, then reassembled. With the holes already prepared for the bolts, no vinyl decal damage occurred.
IMG_20170819_192223585 (resized).jpg

-Coin door, mounted with carriage bolts.
-Start button, typically recessed - drill a large hole with a forstner or spade bit, then drill the small hole all the way through, preferably with a sacrificial wood backer to prevent damage at the exit hole. Make sure the larger hole is deep enough that the start button will not stick out past the cabinet face. Otherwise it is very easy to accidentally trigger it if you lean against the machine.
-Plunger, launch button, or clever game-specific launch handle. Bear in mind that if using a traditional plunger, the playfield mounting depth matters very much. Copy measurements and mount hardware, or if this is not an option, perhaps consider cutting this one hole after you've dropped your whitewood into the cabinet and had a chance to measure it.
-Flipper buttons, recessed the same as the start button but probably with different sized holes.
-Possibly additional buttons for magna-saves or other extended gameplay features.
-Playfield hinge mounts. You might as well decide NOW how you plan on hinging or supporting your playfield, as your decision will affect where you can place playfield mechanisms and where you must have carriage bolt heads sticking through the sides of the cabinet. Make sure you fully understand (even if you're building your own hinge system) where the playfield will sit, how high it will swing up, the footprints of the brackets, and precisely where the holes need to be drilled in the side of the cabinet. If you're basing your cabinet design on an existing game, this is easy - copy it. If you're doing a playfield retheme, half the work is done. If you're like me and you built everything including the frickin hinges, measure it half a dozen times so you don't screw it up.
-A hole for the power button/switch. Make sure it is in a sensible place; test it by FEEL before drilling a hole. Or do as I did, make a blind guess, and then cover that terribly positioned hole with a metal grate and drill a new hole in a better spot.
-Bottom ventilation holes. Get them even and straight if you don't want to feel embarrassed when your machine is folded up on its back.
IMG_20190410_223505807_BURST001 (resized).jpg

-A large hole for the bass speaker. You may want to actually seek out the desired speaker ahead of time so you can measure it and cut the hole appropriately. Or get creative like Scott Danesi and build something custom to make it sound far better.
-Leg bolt holes. A number of people have come up with interesting ways of drilling these nice and straight, and perhaps someone will chime in on this thread with their suggestions. Or do as I did, and eyeball it, ending up with leg bolts that are annoyingly tight.
-Rear ventilation holes.
-Skids. Unless you want the back of your game to look like you scraped it across a few concrete floors (which, most likely, will actually happen), affix wooden runners to the back of the cabinet. Do not attach plastic/rubber feet until cabinet has been painted.
-Mains cord inlet hole. There are a few ways of doing this - a plug receptacle like you would find on a WPC game or a desktop computer, a large hole into a can that holds the plug out of harm's way during moving, or simply cutting a notch at the top of the cabinet for the cord to feed out. Parts can be found with a little shopping around at the usual suppliers.
-Corner joints. If you're building a cabinet, you need to be able to join the corners cleanly. High quality router bits are available for the types of joints you need. They can be expensive. Buy them. You won't regret it, and you can use them for professional results in future projects too, even non-pinball-related.

If you started by repurposing an existing cabinet, you probably need little more than a bit of reconditioning and painting. If you paid to have a cabinet built, you might need to drill a few game-specific holes as needed. Otherwise, you're in for a LOT of woodworking. But if you get a lot of satisfaction from having done an entire project yourself, it's worth it. And cheaper.

I won't dump too much into one post. Backbox advice will come next/later. Anyone else is welcome to join in here - there are many ways of building a pin, and my way is only one of them!

100% this. I spent WAY too much building UT. Part of it was a stubborn desire to build it as traditionally as possible (which resulted in $200+ in bulbs and sockets alone when cheaper modern alternatives are available), and another factor was not really knowing all the parts I needed. I placed more than 60 orders, all told, with various suppliers to build this game, and nearly every single [censored] one of them carried a shipping charge. Looking at it in hindsight and realizing I probably spent $400 or more in F*&%ING FREIGHT COSTS makes me see a pretty major part of my project's expenses right there.

These two places have been a huge asset to the homebrew pinball community. A scratch build requires:
-Woodworking
-Playfield design
-Artwork
-Assembly
-Electronic design and manufacture
-Computer programming
-Audio design
(and probably other fields I can't think of right now, shut up it's late)
P-ROC's offerings essentially knock the fifth item right off that list, and MPF pulls a significant amount of weight off the sixth. I would have taken advantage of both, were I not a stubborn tinkerer and ex video game programmer. (Don't look me up...)

This is another really good one. My layout has three prominent lamp inserts in the center of the playfield, and a green arrow insert pointing into the side passage. Guess what features they represent?

C'mon, guess.

NOTHING.

I have ideas how to take advantage of them, but nothing ended up making it into 1.0 code. Perhaps the green "SAVE" could have been a red "JACKPOT" and the center inserts could have been dropped in favor of collected item status lights. Before diving into your final playfield cut, review all the features you've planned and take note if anything is incomplete. This might even save you some money if you decide to drop a feature that requires additional expense. For instance, I bought a side eject assembly that was originally intended to turn the Grillby scoop into a subway ball lock. When I realized I was going to need more solenoid drivers than I had, I dropped the lock idea - but not before I'd ordered the assembly.

Speaking of features, make sure you know early on whether you want a simple game or a complex one. Complicated games take longer to build, require more programming, and typically cost more - however, I feel they have the advantage of being more interesting to play and, at least to me, more worth the time spent building them. 'Course, if you're re-theming, that decision has already been made for you.

Now, as promised, the backbox. There are MANY ways to build a backbox, particularly because the backbox is a place where you can get creative with the woodworking and still have a game that fits reasonably well next to the rest of a lineup. To cite a notable reference: Look at literally anything the Pinball Amigos build!

My first mistake on my own game: Rushing the construction. If you read the first post in this thread, this should not even be a surprise. All the noticeable flaws in Undertale's backbox could have been prevented by slowing down and measuring more carefully.

If you're going original, the sky's the limit, but if you are big into a particular manufacturer's design choices, get some real measurements first and foremost. I liked the backbox style used on Bally games between around 1989 and early 1992 - display in translite cutout, speakers on top, T-molding on the oversized edges - and chose to build a game to match. It looked pretty nice sitting in my basement.

And then I bought an EATPM, and suddenly my replica backbox didn't look so good anymore.
IMG_20190510_200636176 (resized).jpg

The right thing to do would have been one of two choices. Either drop that design and build something in a style I already owned, or contact an actual owner of a game with the desired construction and politely ask for measurements.

But again - if a replica isn't your goal, GO NUTS. Your backbox is the first eye catcher to a player, you do not have managers dictating your construction constraints, and you can build it in whatever size, shape, and color suits your personal tastes. I refer you to pinballrockstar a second time.

Now here's the thing about the backbox. Depending on your choice of electronics, it may or may not be your most convenient place to store them. But if you decide you're not going to need much backbox wiring, do NOT skimp on wire access holes.

Why, you may ask? Why dig a big hole if you're only running a few wires through it?

Because it might not end up that way.

Undertale was originally going to be driven by a Raspberry Pi until it proved underpowered (more on that later). It needed almost nothing in the backbox - the Pi itself, a tiny power cable to a cheapo audio amplifier, and half a dozen wires to drive some lights and the knocker. There was definitely no reason I would ever have to run more wires than that.

I've seen enough hen--uh, homebrew - to know where this is going.
IMG_20190310_123551319 (resized).jpg

Basically, the big lesson here is give yourself space! Cut large access holes, don't design playfield structures to a tighter tolerance than you can build, and make sure your electronic brains have MORE memory and input amperage than they need, not less.

To follow the trend of the first post, here are the backbox holes you can expect to drill.
-Wire access hole(s). Make them big enough.
-Bolt holes. Drill them into the cabinet too - preferably at the same time, with the backbox positioned where you want it. You might as well mount the rear latch during this process as well, as it'll at least mark where the screws will go after you paint and reassemble.
-Carriage bolt holes for attaching the sides. Even backboxes with routed corner joints tend to be reinforced with angle irons and carriage bolts for additional support, and really why not. Note that you can probably get a better price deal by purchasing a length of raw angle iron and cutting it for your four corners, rather than purchasing four separate shorter brackets. Plus you have control over where the bolts are placed.
-Carriage bolt holes for attaching the hinges. Whether your backbox sits on pivoting strap hinges like a modern title, or door hinges on a neck like an 80s title, you'll want to make sure those hinges are mounted very securely.
-If you intend to use an insert panel - and you should, as it makes your lighting much more efficient - you will want to mount it firmly. Carriage bolts through the left side of the backbox tend to be the go-to solution. Visualize and/or prototype your insert panel hinging system before final assembly to avoid unpleasant surprises related to binding hinges and smashed light bulbs. I got very lucky in this particular situation.
-Your backbox should also have vents, generally near the top. Depending on how much processing power your game needs - and what climate it will be sitting in - you may need to install some fans in these vents. Plenty of hours of testing showed that Undertale did not require any fans besides the built-in power supply and CPU fans.

There is probably more backbox advice to be offered, and I will post it when I think of it!

Very fathomable for me. My RPi3B chugged pretty hard as soon as I tried to push full-screen sprite animation via the Allegro graphics library. No MPF in my project - the entire game code, multimedia aside, was written in raw C++ that ran smooth as silk on the desktop motherboard that eventually powered my game.

Now, granted: I don't know for sure that the Pi was "too slow" for the job. If you could play a realtime 3D game, in software mode, on a PC from 1994, then one of the latest and greatest SBCs really ought to be able to draw static images at 30FPS.

Given that people play HD video on a Pi 3, I do wonder if perhaps the Allegro library was simply not optimized for the hardware. The Pi runs on an ARM processor that is not directly compatible with binaries compiled for x86 or 64. I will be looking into other multimedia libraries for my next build, just in case - I have some past experience with the Irrlicht engine and will have to see if I can get better results. Shader support is nice too.

Yes. My experience was that having the full power of a desktop mobo made the debugging and development process a LOT easier, so I fully agree with this statement. The Pi was also throwing me undervoltage warnings (possibly due to the three dev boards plugged into it), which might have contributed to the poor realtime performance.

I will definitely be looking into a smaller controller for my next build; if I can get the desired graphics working on my Pi3 or a Tinkerboard, I'll go with that - $30 for a computer is a lot cheaper than $200+. Now that I've developed a working game program, I can port it over fairly easily.

Yup. I had no issues with this particular situation on my build, but here's a little addition - you also need wire clearance for backbox hinging. A standard WPC hinge will pull the backbox about a foot away from the cabinet. Your bb->cab wiring might fit nice and tight and tidy with the game assembled, but you might be in for a rude awakening when it comes time to move the game, and you lower the backbox and RIIIIIP.

I had a very close call with backbox wire length the first time around, so when I had to pull my wires out and redo the game, I made sure to allow plenty of clearance the second time around.

This is why the use of simplified programming tools (high level scripting languages and frameworks) gives me pause for concern. Sound cues, and the background functions that a pinball machine would reasonably use, are all fairly lightweight operations that a computer from 1986 had no problem with... think System 11. I don't know how efficient the popular homebrew programming tools are, but I wrote my entire game program from scratch in C++ to avoid any such question. Repeated benchmarking tests proved that my only bottleneck was the full-screen graphics. Had I found a more efficient graphics engine or dropped the HD monitor entirely, my RPi3B would have been more than adequate to run the game.

The problem with avoiding a script library and writing code in raw C/C++ is that you have to literally reinvent the wheel and come up with your own framework. Right down to background timers and light blink patterns. So unless you get great pleasure from low-level software programming, a simple script-based engine and a more powerful computer is your preferred choice.

Ooh, I'm liking this. I will have to check it out.

Thanks for the technical details on MPF! Seems to further reinforce the fact that it's a very well designed system. Didn't know that about the P-Roc library - I suppose that scores another point for writing a custom hardware interface.

Off hand, I don't know if anyone has, but it's certainly possible. I wouldn't be surprised if MPF supports DMD output, I'm sure jabdoa can confirm or deny. For more advanced builders, there's reverse engineering the standard DMD board (plasma or XPin), or constructing a color DMD out of two of these panels:
https://www.adafruit.com/product/2279

I believe there are one or more pinball-ready displays made with this product already, if you don't care to design mounting and drive hardware for these panels from scratch.

As far as animations, you don't need any particular special program to do it. A dot matrix animation is no more than a low-resolution (and usually low-color) string of bitmap images, usually combined with text that could likely be drawn just the same as on an LCD. Exactly how you would create your dot art, and how you would program it to be displayed, would vary based on your choice of computer and hardware.

To me, an advantage to using a DMD instead of an LCD is because the far reduced number of pixels takes a lot less processing power, and your entire video output could be handled by a capable microcontroller. I was giving serious consideration to controlling my entire next game (Volcano Blast) entirely using microcontrollers and a DMD. Ended up going with PC/LCD instead due to the visual clarity and audio capabilities, but the bare metal option is not off the table yet as far as future projects.